Atomic-bomb survivors, certain groups of patients exposed to radiation for medical purposes, and some groups of radiation workers have shown dose-dependent increases in the incidence of certain types of cancer. The induced cancers have not appeared until years after exposure, however, and they have shown no distinguishing features by which they can be identified individually as having resulted from radiation, as opposed to some other cause. With few exceptions, moreover, the incidence of cancer has not been increased detectably by doses of less than 0.01 Sv.

Because the carcinogenic effects of radiation have not been documented over a wide enough range of doses and dose rates to define the shape of the dose-incidence curve precisely, the risk of radiation-induced cancer at low levels of exposure can be estimated only by extrapolation from observations at higher dose levels, based on assumptions about the relation between cancer incidence and dose. For most types of cancer, information about the dose-incidence relationship is rather meagre. The most extensive data available are for leukemia and cancer of the female breast.

The overall incidence of all forms of leukemia other than the chronic lymphatic type has been observed to increase roughly in proportion to dose during the first 25 years after irradiation. Different types of leukemia, however, vary in the magnitude of the radiation-induced increase for a given dose, the age at which irradiation occurs, and the time after exposure. The total excess of all types besides chronic lymphatic leukemia, averaged over all ages, amounts to approximately one to three additional cases of leukemia per year per 10,000 persons at risk per sievert to the bone marrow.

Cancer of the female breast also appears to increase in incidence in proportion to the radiation dose. Furthermore, the magnitude of the increase for a given dose appears to be essentially the same in women whose breasts were irradiated in a single, brief exposure (e.g., atomic-bomb survivors), as in those who were irradiated over a period of years (e.g., patients subjected to multiple fluoroscopic examinations of the chest or workers assigned to coating watch and clock dials with paint containing radium), implying that even small exposures widely separated in time exert carcinogenic effects on the breast that are fully additive and cumulative. Although susceptibility decreases sharply with age at the time of irradiation, the excess of breast cancer averaged over all ages amounts to three to six cases per 10,000 women per sievert each year.

Additional evidence that carcinogenic effects can be produced by a relatively small dose of radiation is provided by the increase in the incidence of thyroid tumours that has been observed to result from a dose of 0.06–2.0 Gy of X rays delivered to the thyroid gland during infancy or childhood, and by the association between prenatal diagnostic X irradiation and childhood leukemia. The latter association implies that exposure to as little as 10–50 mGy of X radiation during intrauterine development may increase the subsequent risk of leukemia in the exposed child by as much as 40–50 percent.

Although some, but not all, other types of cancer have been observed to occur with greater frequency in irradiated populations (Table 12), the data do not suffice to indicate whether the risks extend to low doses. It is apparent, however, that the dose-incidence relationship varies from one type of cancer to another. From the existing evidence, the overall excess of all types of cancer combined may be inferred to approximate 0.6–1.8 cases per 1,000 persons per sievert per year when the whole body is exposed to radiation, beginning two to 10 years after irradiation. This increase corresponds to a cumulative lifetime excess of roughly 20–100 additional cases of cancer per 1,000 persons per sievert, or to an 8–40 percent per sievert increase in the natural lifetime risk of cancer.

The above-cited risk estimates imply that no more than 1–3 percent of all cancers in the general population result from natural background ionizing radiation. At the same time, however, the data suggest that up to 20 percent of lung cancers in nonsmokers may be attributable to inhalation of radon and other naturally occurring radionuclides present in air.

Shortening of the life span

Laboratory animals whose entire bodies are exposed to radiation in the first half of life suffer a reduction in longevity that increases in magnitude with increasing dose. This effect was mistakenly interpreted by early investigators as a manifestation of accelerated or premature aging. The shortening of life in irradiated animals, however, has since been observed to be attributable largely, if not entirely, to the induction of benign and malignant growths. In keeping with this observation is the finding that mortality from diseases other than cancer has not been increased detectably by irradiation among atomic-bomb survivors.